The native snake venom disintegrin bitistatin binds with high affinity to the glycoprotcin IIb/IIIa receptor on activated platelets. When radiolabeled with 123I, bitistatin has been found to be significantly better than related compounds for rapid imaging of experimental pulmonary emboli (PE) and deep venous thrombi (DVT). The reasons for this difference remain to be determined. The overall goal of this project is to optimize a platelet-targeting radiopharmaceutical for imaging PE and DVT, using bitistatin as a basis. First, radioiodinated bitistatin will be more fully characterized in vitro in order to understand its qualities for successful thrombus imaging. The binding of labeled bitistatih to human and canine platelets (both nonstimulated and stimulated) will be assessed. Stability and distribution in whole blood will also be studied. Second, bitistatin will be radiolabeled with Tc-99m using bifunctional chelating agents. The behavior of Tc-99m-bitistatin will be compared with I-125-bitistatin with respect to binding to platelets and other in vitro characteristics. The ability of Tc-99m-bitistatin to image DVT and PE will then be evaluated in an animal model. Image quality and tissue distribution of the Tc-99m labeled bitistatin will be determined and compared with iodinated bitistatin. Third, the possibility of designing smaller peptides which mimic the essential features of bitistatin will be explored, because such peptides could be produced by chemical synthesis. This will be accomplished with the aid of molecular modeling to examine the structural features of bitistatin, in order to determine what unique structural features are present in bitistatin but which are absent in other disintegrrns. The characteristics of both the Arg-Gly-Asp-containing primary binding site and an amino terminal domain, which is unique to bitistatin, will be studied. Disulfideconstrained cyclic peptides will be modeled, and the optimal sizes of loop structures will be tested. The distribution of charged residues, and of hydrophobic residues, will also be considered. Sites for radiolabelin g with 123I or 99mTc will be planned by modeling so that the labeling will not interfere with the binding surface. After the best-fitting peptides have been synthesized and radiolabeled, the behavior of the labeled peptides will be assessed in vitro and compared with labeled native bitistatin. The ability of radiolabeled synthetic peptides to image DVT and PE will then be evaluated in an animal model. Image quality and tissue distribution of the labeled peptides will be determined and compared with labeled native bitistatin. This study should increase our understanding of the in vitro characteristics which are important for imaging thrombotic lesions with platelet-directed radiotracers. Furthermore, the study will identify one or more radiolabeled peptides which have potential clinical utility for rapid, direct imaging of DVT and PE.